Research and System Design of Intelligent License Plate Recognition Algorithm

Abstract

This paper studies the problem of license plate recognition algorithm and system design. The key frame with moving vehicles was extracted by the improved three-frame difference and background subtraction method, then the key frames were pretreated by gray edge detection and morphology, achieved license plate location by the image pixel statistics, and license plate tilt correction, character segmentation based on license plate characters projection method, finally the character feature was extracted using BP neural network algorithm, and output character recognition result, complete license plate recognition. By identifying 100 groups of vehicle videos and images, the license plate recognition rate of the method is over 99%. Based on Open CV, the license plate recognition system is developed to realize the visualization of license plate recognition process.